
# importing required library for developing model and for doing data analysis
import pandas as pd
import glob
import seaborn as sns
import numpy as np
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt  
sns.set(color_codes=True) # adds a nice background to the graphs
%matplotlib inline 
from sklearn.preprocessing import MinMaxScaler
from scipy.stats import zscore
from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import accuracy_score, confusion_matrix
from sklearn.linear_model import LogisticRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.svm import SVC
from sklearn.metrics import classification_report
from sklearn import model_selection
from sklearn.preprocessing import LabelEncoder


filepath = r'.' #assuming the files are in same directory - as this notebook file is
fileslist = glob.glob(filepath + "/Part1*.csv")

li = []

print("shape and size of each file :: \n")
#importing all datasets, exploring each dataset shape and size
for file in fileslist:
    df = pd.read_csv(file, index_col=None, header=0)
    print(file , " is of size :: ", df.size)
    print(file , " is having  ", df.shape[0], "rows,  and ", df.shape[1], " Columns \n")
    li.append(df)

df_concatenated = pd.concat(li, axis=0, ignore_index=True)
#Merge all datasets onto one

shape and size of each file :: 

.\Part1 - Normal.csv  is of size ::  700
.\Part1 - Normal.csv  is having   100 rows,  and  7  Columns 

.\Part1 - Type_H.csv  is of size ::  420
.\Part1 - Type_H.csv  is having   60 rows,  and  7  Columns 

.\Part1 - Type_S.csv  is of size ::  1050
.\Part1 - Type_S.csv  is having   150 rows,  and  7  Columns


#concatenated DataFrame
print("Final dataframe shape :: " ,df_concatenated.shape)
print("Final dataframe size",df_concatenated.size)

Final dataframe shape ::  (310, 7)
Final dataframe size 2170


df_spondylolisthesis  = df_concatenated
df_spondylolisthesis.dtypes

P_incidence    float64
P_tilt         float64
L_angle        float64
S_slope        float64
P_radius       float64
S_Degree       float64
Class           object
dtype: object


#check if there are any null values
df_spondylolisthesis.isnull().values.any()

False


df_spondylolisthesis.sample(15)


df_spondylolisthesis["Class"].unique()

array(['Normal', 'Nrmal', 'Type_H', 'type_h', 'Type_S', 'tp_s'],
      dtype=object)


df_spondylolisthesis.value_counts(df_spondylolisthesis["Class"]).plot(kind="bar");


df_spondylolisthesis.loc[df_spondylolisthesis['Class']=='tp_s','Class']='Type_S'
df_spondylolisthesis.loc[df_spondylolisthesis['Class']=='Nrmal','Class']='Normal'
df_spondylolisthesis.loc[df_spondylolisthesis['Class']=='type_h','Class']='Type_H'


df_spondylolisthesis['Class']=df_spondylolisthesis['Class'].astype('category') #changing to category datatype


df_spondylolisthesis.value_counts(df_spondylolisthesis["Class"]).plot(kind="bar");


df_spondylolisthesis['Class'].nunique()

3


df_spondylolisthesis.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 310 entries, 0 to 309
Data columns (total 7 columns):
 #   Column       Non-Null Count  Dtype   
---  ------       --------------  -----   
 0   P_incidence  310 non-null    float64 
 1   P_tilt       310 non-null    float64 
 2   L_angle      310 non-null    float64 
 3   S_slope      310 non-null    float64 
 4   P_radius     310 non-null    float64 
 5   S_Degree     310 non-null    float64 
 6   Class        310 non-null    category
dtypes: category(1), float64(6)
memory usage: 15.1 KB


#Checking data using describe
df_spondylolisthesis.describe()


import warnings
warnings.simplefilter(action='ignore', category=UserWarning)
warnings.simplefilter(action='ignore', category=FutureWarning)

fig, ax = plt.subplots(3,2, figsize=(18,18))
listOfColumns = ['P_incidence',
 'P_tilt',
 'L_angle',
 'S_slope',
 'P_radius',
 'S_Degree']
len_listOfColumns = len(listOfColumns)
index = 0

for j in range(len(ax)):
    for i in range(len(ax[j])):
        if(index < len_listOfColumns):
            ax[j][i] = sns.distplot(df_spondylolisthesis[listOfColumns[index]],bins=20,rug=True,ax = ax[j][i]);            
            index = index + 1 
fig.show()


import warnings
warnings.simplefilter(action='ignore', category=UserWarning)
warnings.simplefilter(action='ignore', category=FutureWarning)

fig, ax = plt.subplots(3,2, figsize=(18,18))
listOfColumns = ['P_incidence',
 'P_tilt',
 'L_angle',
 'S_slope',
 'P_radius',
 'S_Degree']
len_listOfColumns = len(listOfColumns)
index = 0

for j in range(len(ax)):
    for i in range(len(ax[j])):
        if(index < len_listOfColumns):
            ax[j][i] = sns.boxplot(df_spondylolisthesis[listOfColumns[index]],ax = ax[j][i]);      
            index = index + 1 
fig.show()


listofcolumns = list(df_spondylolisthesis)
for indexOfList in  listofcolumns[0:-1]:
    quant25 = np.percentile(df_spondylolisthesis[indexOfList],25)
    quant75 = np.percentile(df_spondylolisthesis[indexOfList],75)
    Threshold=(quant75-quant25)*1.5
    lower=quant25-Threshold
    upper=quant75+Threshold
    Outliers=[i for i in df_spondylolisthesis[indexOfList] if i < lower or i > upper]
    print('{} Total Number of outliers in {} : {}'.format('\033[1m',indexOfList,len(Outliers)))

 Total Number of outliers in P_incidence : 3
 Total Number of outliers in P_tilt : 13
 Total Number of outliers in L_angle : 1
 Total Number of outliers in S_slope : 1
 Total Number of outliers in P_radius : 11
 Total Number of outliers in S_Degree : 10


import warnings
warnings.simplefilter(action='ignore', category=UserWarning)

fig, ax = plt.subplots(3,2, figsize=(18,18))
listOfColumns = ['P_incidence',
 'P_tilt',
 'L_angle',
 'S_slope',
 'P_radius',
 'S_Degree']
len_listOfColumns = len(listOfColumns)
index = 0

for j in range(len(ax)):
    for i in range(len(ax[j])):
        if(index < len_listOfColumns):
            ax[j][i] = sns.swarmplot(x=df_spondylolisthesis[listOfColumns[index]],y=df_spondylolisthesis['Class'],ax = ax[j][i]) 
            index = index + 1 
fig.show()


import warnings
warnings.simplefilter(action='ignore', category=UserWarning)

fig, ax = plt.subplots(3,2, figsize=(18,18))
listOfColumns = ['P_incidence',
 'P_tilt',
 'L_angle',
 'S_slope',
 'P_radius',
 'S_Degree']
len_listOfColumns = len(listOfColumns)
index = 0

for j in range(len(ax)):
    for i in range(len(ax[j])):
        if(index < len_listOfColumns):
            ax[j][i] = sns.boxplot(x=df_spondylolisthesis[listOfColumns[index]],y=df_spondylolisthesis['Class'],ax = ax[j][i]) 
            index = index + 1 
fig.show()


df_spondylolisthesis.corr()


plt.figure(dpi = 100,figsize= (7,5))
sns.heatmap(df_spondylolisthesis.corr(), fmt = ".2f",annot=True,lw=1,cmap = 'coolwarm')
plt.yticks(rotation = 0)
plt.xticks(rotation = 90)
plt.title('Correlation Heatmap')
plt.show()


sns.pairplot(df_spondylolisthesis,diag_kind='kde',hue='Class');


plt.figure(figsize=(8,6))
sns.countplot('Class',data=df_spondylolisthesis).set_title('Response Variable bar/count Chart')
plt.show()


plt.figure(figsize=(8,6))
df_spondylolisthesis['Class'].value_counts().plot.pie(autopct='%1.1f%%').set_title('Response Variable pie Graph')
plt.show()


class_summary=df_spondylolisthesis.groupby('Class') 
class_summary.mean().reset_index()


#Use label encoder to convert categorical variables to one-hot encoding (dummy variables)
le=LabelEncoder()
df_spondylolisthesis['Class']=le.fit_transform(df_spondylolisthesis['Class'])


df_spondylolisthesis['Class'].value_counts()

2    150
0    100
1     60
Name: Class, dtype: int64


df_spondylolisthesis['Class']=df_spondylolisthesis['Class'].astype('category') #changing datatype to category.


list_of_df_col=['P_incidence','P_tilt','L_angle','S_slope','P_radius','S_Degree']

for i in range(5):
    for col in list_of_df_col:
        quant25 = np.percentile(df_spondylolisthesis[col],25)
        quant75 = np.percentile(df_spondylolisthesis[col],75)
        Threshold=(quant75-quant25)*1.5
        lower=quant25-Threshold
        upper=quant75+Threshold
     
        #Considering mean of a column, but here the values are between  lower and upper cutoff
        df_include = df_spondylolisthesis.loc[(df_spondylolisthesis[col] >= lower) & (df_spondylolisthesis[col] <= upper)]
        mean=int(df_include[col].mean())
       
        #imputing outliers with mean
        df_spondylolisthesis[col]=np.where(df_spondylolisthesis[col]>upper,mean,df_spondylolisthesis[col])
        df_spondylolisthesis[col]=np.where(df_spondylolisthesis[col]<lower,mean,df_spondylolisthesis[col])


df_spondylolisthesis.describe()


listofcolumns = list(df_spondylolisthesis)
for ind in  listofcolumns[0:-1]:
    quant25 = np.percentile(df_spondylolisthesis[ind],25)
    quant75 = np.percentile(df_spondylolisthesis[ind],75)
    Threshold=(quant75-quant25)*1.5
    lower=quant25-Threshold
    upper=quant75+Threshold
    Outliers=[i for i in df_spondylolisthesis[ind] if i < lower or i > upper]
    print('{} Total Number of outliers in {} : {}'.format('\033[1m',ind,len(Outliers)))

 Total Number of outliers in P_incidence : 0
 Total Number of outliers in P_tilt : 0
 Total Number of outliers in L_angle : 0
 Total Number of outliers in S_slope : 0
 Total Number of outliers in P_radius : 0
 Total Number of outliers in S_Degree : 0


#Segregate features and class variables
x_data = df_spondylolisthesis[['P_incidence',
 'P_tilt',
 'L_angle',
 'S_slope',
 'P_radius',
 'S_Degree']]
y_data = df_spondylolisthesis['Class']


x_data


#standardization of data
from sklearn import preprocessing
from sklearn.preprocessing import StandardScaler
X_std = pd.DataFrame(StandardScaler().fit_transform(x_data))
X_std.columns = x_data.columns


#Check for target balancing. 
y_data.value_counts().plot(kind="bar");


#split the dataset into training and test datasets
# Transform data into features and target
X = np.array(X_std) 
y = np.array(y_data)

# split into train and test
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=10)


# loading library
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import accuracy_score
from sklearn.metrics import recall_score

# instantiate learning model (k = 3)
knn = KNeighborsClassifier(n_neighbors= 5 , metric = 'euclidean')

# fitting the model
knn.fit(X_train, y_train)


y_pred_train = knn.predict(X_train)
# predict the response
y_pred = knn.predict(X_test)


from sklearn.metrics import accuracy_score
train_accuracy=accuracy_score(y_train, y_pred_train)
print('train Accuracy: ',np.round(train_accuracy,4))

# print classification report
print("classification report for train")
print(classification_report(y_train, y_pred_train))

train Accuracy:  0.8848
classification report for train
              precision    recall  f1-score   support

           0       0.84      0.88      0.86        72
           1       0.79      0.79      0.79        42
           2       0.96      0.93      0.95       103

    accuracy                           0.88       217
   macro avg       0.86      0.86      0.86       217
weighted avg       0.89      0.88      0.89       217


from sklearn.metrics import accuracy_score
train_accuracy=accuracy_score(y_test, y_pred)
print('train Accuracy: ',np.round(train_accuracy,4))

# print classification report
print("classification report for train")
print(classification_report(y_test, y_pred))

train Accuracy:  0.8065
classification report for train
              precision    recall  f1-score   support

           0       0.73      0.68      0.70        28
           1       0.58      0.78      0.67        18
           2       0.98      0.89      0.93        47

    accuracy                           0.81        93
   macro avg       0.76      0.78      0.77        93
weighted avg       0.83      0.81      0.81        93


cm = confusion_matrix(y_test, y_pred, labels=[0, 1, 2])

df_cm = pd.DataFrame(cm, index = [i for i in ["Normal","Type_H","Type_S"]],
                  columns = [i for i in ["Normal","Type_H","Type_S"]])
plt.figure(figsize = (7,5))
sns.heatmap(df_cm, annot=True ,fmt='g')
plt.show()


from sklearn.model_selection import GridSearchCV
from sklearn.neighbors import KNeighborsClassifier
#making the instance
model = KNeighborsClassifier(n_jobs=-1)
#Hyper Parameters Set
params = {'n_neighbors':[3,5,7],
          'leaf_size':[1,3,5,7],
          'weights':['uniform', 'distance'],
          'algorithm':['auto', 'ball_tree','kd_tree','brute'],
          'n_jobs':[-1],
          'p':[1,2,3,4]}
#Making models with hyper parameters sets
model1 = GridSearchCV(model, param_grid=params, n_jobs=1)
#Learning
model1.fit(X_train,y_train)
#The best hyper parameters set
print("Best Hyper Parameters:\n",model1.best_params_)
#Prediction
prediction=model1.predict(X_test)
#importing the metrics module
from sklearn import metrics
#evaluation(Accuracy)
print("Accuracy:",metrics.accuracy_score(prediction,y_test))
# print(recall_score(y_test, prediction))
#evaluation(Confusion Metrix)

Best Hyper Parameters:
 {'algorithm': 'auto', 'leaf_size': 1, 'n_jobs': -1, 'n_neighbors': 3, 'p': 3, 'weights': 'distance'}
Accuracy: 0.7741935483870968


# subsetting just the odd ones, it is best to set number of neighbours to odd numbers
# neighbors = list(filter(lambda x: x % 2 != 0, list(range(1,20))))
neighbors = list(range(1, 50, 2))

# empty list that will hold accuracy scores which will hold the required accuracy scores
ac_scores = []

# perform accuracy metrics for values from 1,3,5....19
for k in neighbors:
    knn = KNeighborsClassifier(n_neighbors=k)
    knn.fit(X_train, y_train)
    # predict the response
    y_pred = knn.predict(X_test)
    # evaluate accuracy
    scores = accuracy_score(y_test, y_pred)
    ac_scores.append(scores)

# changing to misclassification error
MSE = [1 - x for x in ac_scores]

# determining best k
optimal_k = neighbors[MSE.index(min(MSE))]
print("The optimal number of neighbors is %d" % optimal_k)

The optimal number of neighbors is 5


import matplotlib.pyplot as plt
# plot misclassification error vs k
plt.plot(neighbors, MSE)
plt.xlabel('Number of Neighbors K')
plt.ylabel('Misclassification Error')
plt.show()


train_score=[]
test_score=[]
for k in range(1,51):
    KNN = KNeighborsClassifier(n_neighbors= k , metric = 'euclidean' ) 
    KNN.fit(X_train, y_train)
    train_score.append(KNN.score(X_train, y_train))
    test_score.append(KNN.score(X_test, y_test))
    
plt.plot(range(1,51),train_score)
plt.plot(range(1,51),test_score)
plt.show()


from sklearn.model_selection import cross_val_score
# creating odd list of K for KNN
neighbors = list(range(1, 50, 2))

# empty list that will hold cv scores
cv_scores = []

# perform 10-fold cross validation
for k in neighbors:
    knn = KNeighborsClassifier(n_neighbors=k)
    scores = cross_val_score(knn, X_train, y_train, cv=15, scoring='accuracy')
    cv_scores.append(scores.mean())
    
# changing to misclassification error
mse = [1 - x for x in cv_scores]

# determining best k
optimal_k = neighbors[mse.index(min(mse))]
print("The optimal number of neighbors is {}".format(optimal_k))

# plot misclassification error vs k
plt.plot(neighbors, mse)
plt.xlabel("Number of Neighbors K")
plt.ylabel("Misclassification Error")
plt.show()

The optimal number of neighbors is 11


# Pandas and Numpy libraries
import pandas as pd
import numpy as np
# For preprocessing the data
from sklearn import preprocessing
# To split the dataset into train and test dataset

from sklearn.model_selection import train_test_split
# To model the Gaussian Navie Bayes classifier
from sklearn.naive_bayes import GaussianNB

#standardization of data
from sklearn import preprocessing
from sklearn.preprocessing import StandardScaler
from sklearn import metrics
from sklearn.metrics import average_precision_score, confusion_matrix, accuracy_score, classification_report

import matplotlib.pyplot as plt       # matplotlib.pyplot plots data
%matplotlib inline 
import seaborn as sns

sns.set(color_codes=True) # adds a nice background to the graphs
%matplotlib inline 

import warnings
warnings.filterwarnings('ignore')
pd.options.display.max_columns = None
pd.options.display.float_format = '{:.2f}'.format
pd.options.display.max_rows = None


#Importing dataset
part1BAFDf = pd.read_csv("./Part2 - Data1.csv")
part2BAFDf = pd.read_csv("./Part2 -Data2.csv")


#Checking shape and size of dataset
print("size of ./Part2 - Data1.csv",part1BAFDf.size)
print("Number of rows in ./Part2 - Data1.csv",part1BAFDf.shape[0], "Number of columns in ./Part2 - Data1.csv",part1BAFDf.shape[1])
print("Number of rows in ./Part2 - Data1.csv",part2BAFDf.shape[0], "Number of columns in ./Part2 - Data1.csv",part2BAFDf.shape[1])

size of ./Part2 - Data1.csv 40000
Number of rows in ./Part2 - Data1.csv 5000 Number of columns in ./Part2 - Data1.csv 8
Number of rows in ./Part2 - Data1.csv 5000 Number of columns in ./Part2 - Data1.csv 7


#merging all dataset
banking_Finance_dataset_Raw = pd.merge(part1BAFDf, part2BAFDf, how ='inner', on ='ID')


#creating a copy of dataset - which could be useful in EDA
banking_Finance_dataset = banking_Finance_dataset_Raw.copy()


print("size of final data set :: " , banking_Finance_dataset.size)
print("No of rows in final data set :: " , banking_Finance_dataset.shape[0])
print("No of columns in final data set :: " , banking_Finance_dataset.shape[1])

size of final data set ::  70000
No of rows in final data set ::  5000
No of columns in final data set ::  14


banking_Finance_dataset.sample(10)


banking_Finance_dataset.columns

Index(['ID', 'Age', 'CustomerSince', 'HighestSpend', 'ZipCode', 'HiddenScore',
       'MonthlyAverageSpend', 'Level', 'Mortgage', 'Security',
       'FixedDepositAccount', 'InternetBanking', 'CreditCard', 'LoanOnCard'],
      dtype='object')


banking_Finance_dataset.describe()


banking_Finance_dataset.dtypes

ID                       int64
Age                      int64
CustomerSince            int64
HighestSpend             int64
ZipCode                  int64
HiddenScore              int64
MonthlyAverageSpend    float64
Level                    int64
Mortgage                 int64
Security                 int64
FixedDepositAccount      int64
InternetBanking          int64
CreditCard               int64
LoanOnCard             float64
dtype: object


for i in ['HiddenScore','Level' , 'Security' , 'FixedDepositAccount' , 'InternetBanking'  , 'CreditCard' , 'LoanOnCard']:
    print(i , banking_Finance_dataset[i].unique());
    print(i , len(banking_Finance_dataset[i].unique()));

HiddenScore [4 3 1 2]
HiddenScore 4
Level [1 2 3]
Level 3
Security [1 0]
Security 2
FixedDepositAccount [0 1]
FixedDepositAccount 2
InternetBanking [0 1]
InternetBanking 2
CreditCard [0 1]
CreditCard 2
LoanOnCard [nan  1.  0.]
LoanOnCard 3


for i in ['HiddenScore','Level' , 'Security' , 'FixedDepositAccount' , 'InternetBanking'  , 'CreditCard' , 'LoanOnCard']:
    banking_Finance_dataset[i]=banking_Finance_dataset[i].astype('category') #changing datatype to category.


banking_Finance_dataset.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 5000 entries, 0 to 4999
Data columns (total 14 columns):
 #   Column               Non-Null Count  Dtype   
---  ------               --------------  -----   
 0   ID                   5000 non-null   int64   
 1   Age                  5000 non-null   int64   
 2   CustomerSince        5000 non-null   int64   
 3   HighestSpend         5000 non-null   int64   
 4   ZipCode              5000 non-null   int64   
 5   HiddenScore          5000 non-null   category
 6   MonthlyAverageSpend  5000 non-null   float64 
 7   Level                5000 non-null   category
 8   Mortgage             5000 non-null   int64   
 9   Security             5000 non-null   category
 10  FixedDepositAccount  5000 non-null   category
 11  InternetBanking      5000 non-null   category
 12  CreditCard           5000 non-null   category
 13  LoanOnCard           4980 non-null   category
dtypes: category(7), float64(1), int64(6)
memory usage: 507.4 KB


banking_Finance_dataset.isnull().values.any() # If there are any null values in data set

True


banking_Finance_dataset.isnull().values.sum()

20


banking_Finance_dataset = banking_Finance_dataset.dropna(axis = 0, how ='any')


banking_Finance_dataset.isnull().values.sum()

0


#check final count in the dataset
print("number of records after removing outlier" ,banking_Finance_dataset.shape)

number of records after removing outlier (4980, 14)


banking_Finance_dataset['LoanOnCard'].unique()

[1.00, 0.00]
Categories (2, float64): [1.00, 0.00]


banking_Finance_dataset['LoanOnCard'].value_counts(normalize=True)

0.0   0.90
1.0   0.10
Name: LoanOnCard, dtype: float64


ax = sns.countplot(x="LoanOnCard", hue="CreditCard", data=banking_Finance_dataset)


banking_Finance_dataset.describe()


banking_Finance_dataset.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 4980 entries, 9 to 4999
Data columns (total 14 columns):
 #   Column               Non-Null Count  Dtype   
---  ------               --------------  -----   
 0   ID                   4980 non-null   int64   
 1   Age                  4980 non-null   int64   
 2   CustomerSince        4980 non-null   int64   
 3   HighestSpend         4980 non-null   int64   
 4   ZipCode              4980 non-null   int64   
 5   HiddenScore          4980 non-null   category
 6   MonthlyAverageSpend  4980 non-null   float64 
 7   Level                4980 non-null   category
 8   Mortgage             4980 non-null   int64   
 9   Security             4980 non-null   category
 10  FixedDepositAccount  4980 non-null   category
 11  InternetBanking      4980 non-null   category
 12  CreditCard           4980 non-null   category
 13  LoanOnCard           4980 non-null   category
dtypes: category(7), float64(1), int64(6)
memory usage: 506.1 KB


import warnings
warnings.simplefilter(action='ignore', category=FutureWarning)

fig, ax = plt.subplots(3,2, figsize=(18,18))
listOfColumns = ['Age',
 'CustomerSince',
 'HighestSpend',
 'MonthlyAverageSpend',
 'Mortgage']
len_listOfColumns = len(listOfColumns)
index = 0

for j in range(len(ax)):
    for i in range(len(ax[j])):
        if(index < len_listOfColumns):
            ax[j][i] = sns.boxplot(banking_Finance_dataset[listOfColumns[index]],ax = ax[j][i]);      
            #ax[j][i] = sns.violinplot(concatenated_df[listOfColumns[index]],ax = ax[j][i]);      
            index = index + 1 
fig.show()


import warnings
warnings.simplefilter(action='ignore', category=FutureWarning)

fig, ax = plt.subplots(3,2, figsize=(18,18))
listOfColumns = ['Age',
 'CustomerSince',
 'HighestSpend',
 'MonthlyAverageSpend',
 'Mortgage']
len_listOfColumns = len(listOfColumns)
index = 0

for j in range(len(ax)):
    for i in range(len(ax[j])):
        if(index < len_listOfColumns):
            ax[j][i] = sns.distplot(banking_Finance_dataset[listOfColumns[index]],ax = ax[j][i]);      
            #ax[j][i] = sns.violinplot(concatenated_df[listOfColumns[index]],ax = ax[j][i]);      
            index = index + 1 
fig.show()


listofcolumns = ['Age','CustomerSince','HighestSpend','MonthlyAverageSpend','Mortgage']
for ind in  listofcolumns:
    quant25 = np.percentile(banking_Finance_dataset[ind],25)
    quant75 = np.percentile(banking_Finance_dataset[ind],75)
    Threshold=(quant75-quant25)*1.5
    lower=quant25-Threshold
    upper=quant75+Threshold
    Outliers=[i for i in banking_Finance_dataset[ind] if i < lower or i > upper]
    print('{} Total Number of outliers in {} : {}'.format('\033[1m',ind,len(Outliers)))

 Total Number of outliers in Age : 0
 Total Number of outliers in CustomerSince : 0
 Total Number of outliers in HighestSpend : 96
 Total Number of outliers in MonthlyAverageSpend : 324
 Total Number of outliers in Mortgage : 291


import warnings
warnings.simplefilter(action='ignore', category=UserWarning)

fig, ax = plt.subplots(3,2, figsize=(18,18))
listOfColumns = ['Age', 'CustomerSince',
 'HighestSpend','Mortgage',
 'MonthlyAverageSpend']
len_listOfColumns = len(listOfColumns)
index = 0

for j in range(len(ax)):
    for i in range(len(ax[j])):
        if(index < len_listOfColumns):
            ax[j][i] = sns.boxplot(x=banking_Finance_dataset[listOfColumns[index]],y=banking_Finance_dataset['LoanOnCard'],ax = ax[j][i]) 
            index = index + 1 
fig.show()


sns.pairplot(banking_Finance_dataset_Raw,diag_kind='kde',hue='LoanOnCard');


sns.pairplot(banking_Finance_dataset,diag_kind='kde',hue='LoanOnCard');


banking_Finance_dataset.corr()


import seaborn as sns
%matplotlib inline

# calculate the correlation matrix
corr = banking_Finance_dataset.corr()

plt.figure(dpi = 100,figsize=(20,10))
sns.heatmap(corr, fmt = ".2f",annot=True,lw=1,cmap = 'coolwarm')
plt.yticks(rotation = 0)
plt.xticks(rotation = 90)
plt.title('Correlation Heatmap')
plt.show()


import seaborn as sns
%matplotlib inline

plt.figure(dpi = 100,figsize=(20,10))
sns.heatmap(banking_Finance_dataset_Raw.corr(), fmt = ".2f",annot=True,lw=1,cmap = 'coolwarm')
plt.yticks(rotation = 0)
plt.xticks(rotation = 90)
plt.title('Correlation Heatmap')
plt.show()


plt.figure(figsize=(8,6))
sns.countplot('LoanOnCard',data=banking_Finance_dataset_Raw).set_title('Response Variable bar/count Chart')
plt.show()


plt.figure(figsize=(8,6))
banking_Finance_dataset['LoanOnCard'].value_counts().plot.pie(autopct='%1.1f%%').set_title('Response Variable pie Graph')
plt.show()


class_summary=banking_Finance_dataset_Raw[['Age','CustomerSince','HighestSpend','HiddenScore','MonthlyAverageSpend','Level','Mortgage','LoanOnCard']].groupby('LoanOnCard') 
class_summary.mean().reset_index()


#Segregate features and class variables
x_data = banking_Finance_dataset[[  
 'Age',
 'HighestSpend',
 'HiddenScore',
 'MonthlyAverageSpend',
 'Level',
 'Mortgage',
 'Security',
 'FixedDepositAccount',
 'InternetBanking'
]]
y_data = banking_Finance_dataset['LoanOnCard']


x_data['Mortgage'].value_counts()

0      3447
98       17
91       16
83       16
89       16
119      16
103      16
78       15
102      15
90       15
94       14
87       14
101      14
131      14
144      13
109      13
118      13
81       13
116      13
112      13
104      13
106      13
185      12
97       12
120      12
100      12
121      12
76       12
184      11
95       11
79       11
111      11
86       11
153      11
115      11
137      11
84       11
158      11
82       10
135      10
108      10
151      10
142      10
161      10
117      10
113      10
123       9
167       9
128       9
146       9
221       9
166       9
148       9
88        9
149       9
159       9
110       8
75        8
114       8
122       8
138       8
170       8
194       8
218       8
147       8
169       8
105       8
132       8
157       8
164       8
207       8
205       8
140       8
230       7
129       7
219       7
204       7
193       7
154       7
124       7
85        7
96        7
229       7
174       7
127       7
80        7
249       7
196       7
125       6
240       6
126       6
130       6
134       6
136       6
236       6
145       6
150       6
182       6
141       6
251       6
163       6
192       6
187       6
107       6
99        6
212       5
220       5
198       5
171       5
178       5
256       5
217       5
209       5
294       5
155       5
172       5
301       5
180       5
227       5
188       5
272       5
307       5
156       5
297       4
93        4
162       4
241       4
181       4
190       4
245       4
189       4
77        4
247       4
211       4
232       4
199       4
224       4
244       4
342       4
152       4
179       4
175       4
203       4
231       4
139       4
200       4
92        4
223       3
352       3
328       3
308       3
177       3
197       3
239       3
165       3
215       3
226       3
255       3
160       3
380       3
400       3
341       3
143       3
333       3
329       3
325       3
428       3
257       3
305       3
176       3
168       3
213       3
263       3
248       3
310       3
214       3
260       3
264       3
238       3
242       3
186       3
327       3
323       3
270       3
319       3
282       3
285       3
315       3
277       3
133       3
228       3
268       3
275       3
422       3
366       3
216       3
208       3
427       2
243       2
267       2
455       2
271       2
303       2
359       2
299       2
233       2
287       2
225       2
183       2
201       2
259       2
222       2
289       2
312       2
292       2
280       2
276       2
442       2
402       2
394       2
358       2
354       2
330       2
322       2
314       2
306       2
252       2
266       2
262       2
250       2
246       2
234       2
202       2
565       2
304       2
392       2
364       2
293       2
309       2
313       2
357       2
372       2
397       2
368       2
408       2
437       2
449       2
360       1
416       1
432       1
452       1
464       1
508       1
496       1
300       1
412       1
284       1
500       1
524       1
296       1
396       1
344       1
336       1
635       1
612       1
283       1
195       1
191       1
590       1
582       1
550       1
522       1
470       1
466       1
458       1
446       1
410       1
406       1
398       1
382       1
378       1
374       1
334       1
235       1
295       1
318       1
311       1
571       1
567       1
547       1
535       1
483       1
475       1
467       1
431       1
419       1
415       1
403       1
391       1
383       1
355       1
351       1
343       1
331       1
326       1
302       1
173       1
461       1
429       1
421       1
405       1
389       1
385       1
381       1
373       1
361       1
353       1
345       1
337       1
321       1
281       1
273       1
265       1
253       1
237       1
433       1
477       1
298       1
481       1
290       1
286       1
278       1
258       1
210       1
206       1
617       1
601       1
589       1
581       1
587       1
569       1
553       1
541       1
509       1
505       1
485       1
577       1
Name: Mortgage, dtype: int64


x_data['MonthlyAverageSpend'].value_counts()

0.30     240
1.00     229
0.20     204
2.00     188
0.80     187
0.10     183
0.40     177
1.50     176
0.70     169
0.50     163
1.70     158
1.80     151
1.40     136
2.20     129
1.30     127
0.60     117
2.80     110
2.50     107
1.90     106
0.00     106
0.90     105
2.10     100
1.60     100
2.40      91
2.60      87
1.10      82
1.20      66
2.30      57
2.70      56
2.90      54
3.00      53
3.30      45
3.80      43
3.40      39
2.67      36
4.00      32
4.50      29
3.90      27
3.60      27
4.30      26
6.00      26
3.70      25
4.70      24
4.90      22
3.20      22
4.10      22
3.10      20
6.50      18
1.67      18
2.33      18
5.40      18
0.67      18
5.00      18
4.40      17
5.20      16
3.50      15
6.90      14
7.00      14
4.60      14
6.10      14
6.30      13
5.70      13
7.20      13
7.40      13
7.50      12
8.00      12
4.20      11
6.33      10
7.30      10
6.80      10
8.10      10
1.33       9
7.80       9
0.75       9
8.80       9
1.75       9
6.67       9
6.70       9
7.60       9
4.33       9
8.60       8
4.80       7
5.60       7
5.10       6
5.90       5
7.90       4
5.50       4
5.30       4
6.60       4
5.80       3
10.00      3
6.40       3
5.67       2
4.75       2
8.30       2
4.25       2
6.20       2
8.50       2
9.00       2
4.67       1
3.67       1
2.75       1
3.33       1
3.25       1
8.20       1
9.30       1
8.90       1
5.33       1
Name: MonthlyAverageSpend, dtype: int64


x_data['HighestSpend'].value_counts()

44     85
38     84
81     82
41     81
39     81
40     77
42     76
83     74
43     70
45     68
29     66
85     65
35     65
21     65
25     64
22     64
30     63
28     63
84     63
55     61
82     60
64     60
78     60
65     60
32     58
61     57
53     57
80     56
58     55
62     55
31     55
23     54
79     53
59     53
54     52
34     52
60     52
18     52
19     51
49     51
33     51
75     47
24     47
70     47
20     47
52     47
63     46
69     46
74     45
50     44
48     44
73     44
71     43
72     40
51     40
90     38
93     37
91     37
68     35
89     34
113    33
15     33
13     32
14     31
114    30
12     30
92     29
98     28
115    27
94     26
11     26
9      26
112    26
88     26
95     25
141    24
101    24
99     24
122    24
128    24
129    23
145    23
125    23
8      23
10     23
111    22
154    21
121    20
105    20
149    20
104    20
134    20
110    19
130    19
118    19
131    19
140    19
155    19
123    18
103    18
135    18
138    18
180    18
158    18
132    18
119    18
179    17
109    17
120    17
102    16
139    16
108    16
161    16
152    15
133    15
195    15
142    15
164    13
173    13
182    13
191    13
184    12
175    12
170    12
124    12
160    12
183    12
172    11
148    11
165    11
153    11
150    11
190    11
178    10
188    10
162    10
100     9
174     9
171     9
163     9
185     9
143     9
194     8
181     8
168     8
144     7
159     7
169     7
192     6
193     6
201     5
151     4
204     3
200     3
198     3
199     3
203     2
189     2
202     2
205     2
224     1
218     1
Name: HighestSpend, dtype: int64


#standardization of data
from sklearn import preprocessing
from sklearn.preprocessing import StandardScaler
X_std = pd.DataFrame(StandardScaler().fit_transform(x_data))
X_std.columns = x_data.columns


from sklearn.model_selection import train_test_split
print("Number of records standard Data X: ", X_std.shape)
X_tr, X_tt, y_tr, y_tt = train_test_split(X_std, y_data, test_size = 0.30, random_state = 10)
print("Number of records after splitting X train: ", X_tr.shape)
print("Number of records after splitting X test: ", X_tt.shape)
print("Number of records after splitting y train: ", y_tr.shape)
print("Number of records after splitting y test: ", y_tt.shape)

Number of records standard Data X:  (4980, 9)
Number of records after splitting X train:  (3486, 9)
Number of records after splitting X test:  (1494, 9)
Number of records after splitting y train:  (3486,)
Number of records after splitting y test:  (1494,)


y_tt.to_frame().value_counts(normalize=True)

LoanOnCard
0.0          0.90
1.0          0.10
dtype: float64


y_tr.to_frame().value_counts(normalize=True)

LoanOnCard
0.0          0.90
1.0          0.10
dtype: float64


from imblearn.over_sampling import RandomOverSampler
# define oversampling strategy
oversample = RandomOverSampler(sampling_strategy='minority')
# fit and apply the transform
X_over, y_over = oversample.fit_resample(X_tr, y_tr)
print("Number of records oversample Data X: ", X_over.shape)
print("Number of records oversample Data y: ", y_over.shape)
print("Number of records split train Data X: ", X_tr.shape)
print("Number of records split train Data y: ", y_tr.shape)

Number of records oversample Data X:  (6306, 9)
Number of records oversample Data y:  (6306,)
Number of records split train Data X:  (3486, 9)
Number of records split train Data y:  (3486,)


##Design and train a Logistic regression and Naive Bayes classifiers.

# logistic regression object
from sklearn.linear_model import LogisticRegression

#parameter are choosen based on gridsearch output. This is available in two cells below
lr=LogisticRegression(penalty='l2',C=1,solver = 'liblinear',max_iter=300)
# train the model on train set using oversampled train data 
lr.fit(X_over, y_over)

predictions_train = lr.predict(X_over)
predictions_test = lr.predict(X_tt)


from sklearn.metrics import accuracy_score
train_accuracy=accuracy_score(y_over, predictions_train)
print('train Accuracy: ',np.round(train_accuracy,4))

# print classification report
print("classification report for train")
print(classification_report(y_over, predictions_train))

train Accuracy:  0.9028
classification report for train
              precision    recall  f1-score   support

         0.0       0.91      0.90      0.90      3153
         1.0       0.90      0.91      0.90      3153

    accuracy                           0.90      6306
   macro avg       0.90      0.90      0.90      6306
weighted avg       0.90      0.90      0.90      6306


from sklearn.metrics import accuracy_score
test_accuracy=accuracy_score(y_tt, predictions_test)
print('Test Accuracy: ',np.round(test_accuracy,4))

# print classification report
print("classification report for test")
print(classification_report(y_tt, predictions_test))

Test Accuracy:  0.8762
classification report for test
              precision    recall  f1-score   support

         0.0       0.99      0.87      0.93      1347
         1.0       0.44      0.91      0.59       147

    accuracy                           0.88      1494
   macro avg       0.71      0.89      0.76      1494
weighted avg       0.93      0.88      0.89      1494


coef_df = pd.DataFrame(lr.coef_)
coef_df['intercept'] = lr.intercept_
print(coef_df)

      0    1    2    3    4    5     6    7     8  intercept
0 -0.08 2.54 0.64 0.38 1.18 0.04 -0.27 0.75 -0.17      -2.35


cm=metrics.confusion_matrix(y_tt, predictions_test, labels=[1, 0])

df_cm = pd.DataFrame(cm, index = [i for i in ["1","0"]],
                  columns = [i for i in ["Predict 1","Predict 0"]])

plt.figure(figsize = (7,5))
sns.heatmap(df_cm, annot=True, fmt='g');


from sklearn.model_selection import GridSearchCV
from sklearn.linear_model import LogisticRegression
grid={"C":[-1,0,1,10,100], "penalty":["l1","l2"], 'solver': ['newton-cg', 'lbfgs', 'liblinear', 'sag', 'saga']}# l1 lasso l2 ridge
logreg=LogisticRegression()
logreg_cv=GridSearchCV(logreg,grid,cv=10)
logreg_cv.fit(X_over, y_over)

print("tuned hpyerparameters :(best parameters) ",logreg_cv.best_params_)
print("accuracy :",logreg_cv.best_score_)

tuned hpyerparameters :(best parameters)  {'C': 1, 'penalty': 'l2', 'solver': 'liblinear'}
accuracy : 0.9024760395441854


from sklearn.model_selection import cross_val_score

scores = cross_val_score(lr, X_over, y_over, cv=100)
print('Cross-Validation Accuracy Scores', scores)
scores = pd.Series(scores)
scores.min(), scores.mean(), scores.max()

Cross-Validation Accuracy Scores [0.90625    0.90625    0.90625    0.8125     0.921875   0.84375
 0.85714286 0.92063492 0.82539683 0.88888889 0.84126984 0.92063492
 0.92063492 0.93650794 0.95238095 0.88888889 0.92063492 0.92063492
 0.96825397 0.92063492 0.88888889 0.93650794 1.         0.92063492
 0.93650794 0.88888889 0.87301587 0.88888889 0.85714286 0.79365079
 0.95238095 0.9047619  0.88888889 0.88888889 0.92063492 0.92063492
 0.95238095 0.87301587 0.95238095 0.87301587 0.93650794 0.92063492
 0.87301587 0.87301587 0.93650794 0.79365079 0.80952381 0.92063492
 0.88888889 0.93650794 0.88888889 0.82539683 0.9047619  0.98412698
 0.93650794 0.87301587 0.87301587 0.87301587 0.92063492 0.92063492
 0.85714286 0.88888889 0.87301587 0.88888889 0.9047619  0.88888889
 0.93650794 0.87301587 0.88888889 0.87301587 0.82539683 0.93650794
 0.92063492 0.9047619  0.95238095 0.95238095 0.88888889 0.93650794
 0.82539683 0.93650794 0.95238095 0.96825397 0.96825397 0.82539683
 0.9047619  0.88888889 0.88888889 0.9047619  0.93650794 0.96825397
 0.9047619  0.82539683 0.9047619  0.92063492 0.92063492 0.9047619
 0.9047619  0.92063492 0.88888889 0.88888889]

(0.7936507936507936, 0.9012227182539684, 1.0)


clf = GaussianNB()
clf.fit(X_over, y_over)
nb_predictions_test = clf.predict(X_tt)


from sklearn.metrics import accuracy_score
nb_test_accuracy=accuracy_score(y_tt, nb_predictions_test)
print('Test Accuracy: ',np.round(nb_test_accuracy,4))

# print classification report
print("classification report for test")
print(classification_report(y_tt, nb_predictions_test))

Test Accuracy:  0.8568
classification report for test
              precision    recall  f1-score   support

         0.0       0.97      0.86      0.92      1347
         1.0       0.39      0.78      0.52       147

    accuracy                           0.86      1494
   macro avg       0.68      0.82      0.72      1494
weighted avg       0.92      0.86      0.88      1494


#For NB
cm=metrics.confusion_matrix(y_tt, nb_predictions_test, labels=[1, 0])

df_cm = pd.DataFrame(cm, index = [i for i in ["1","0"]],
                  columns = [i for i in ["Predict 1","Predict 0"]])

plt.figure(figsize = (7,5))
sns.heatmap(df_cm, annot=True, fmt='g');


#AUC ROC curve
from sklearn.metrics import roc_auc_score
from sklearn.metrics import roc_curve

logit_roc_auc = roc_auc_score(y_tt, lr.predict_proba(X_tt)[:,1])
fpr, tpr, thresholds = roc_curve(y_tt, lr.predict_proba(X_tt)[:,1])
plt.figure()
plt.plot(fpr, tpr, label='Logistic Regression (area = %0.2f)' % logit_roc_auc)
plt.plot([0, 1], [0, 1],'r--',label='No Skill')
plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05])
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.title('Receiver operating characteristic')
plt.legend(loc="lower right")
plt.savefig('Log_ROC')
plt.show()

print(logit_roc_auc)

0.9540273421915164


#AUC ROC curve
from sklearn.metrics import roc_auc_score
from sklearn.metrics import roc_curve
# roc curve for logistic regression model with optimal threshold
from numpy import sqrt
from numpy import argmax

logit_roc_auc = roc_auc_score(y_tt, lr.predict_proba(X_tt)[:,1])
fpr, tpr, thresholds = roc_curve(y_tt, lr.predict_proba(X_tt)[:,1])
# calculate the g-mean for each threshold
gmeans = sqrt(tpr * (1-fpr))
# locate the index of the largest g-mean
ix = argmax(gmeans)
print('Best Threshold=%f, G-Mean=%.3f' % (thresholds[ix], gmeans[ix]))
plt.figure()
plt.plot(fpr, tpr, label='Logistic Regression (area = %0.2f)' % logit_roc_auc)
plt.plot([0, 1], [0, 1],'r--',label='No Skill')
plt.scatter(fpr[ix], tpr[ix], marker='o', color='black', label='Best')
plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05])
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.title('Receiver operating characteristic')
plt.legend(loc="lower right")
plt.savefig('Log_ROC')
plt.show()

Best Threshold=0.536350, G-Mean=0.899

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree	Class
184	76.326002	42.396204	57.200000	33.929797	124.267007	50.127457	Type_S
124	36.125683	22.758753	29.000000	13.366931	115.577116	-3.237562	Type_H
307	74.094731	18.823727	76.032156	55.271004	128.405731	73.388216	Type_S
253	57.286945	15.149350	64.000000	42.137595	116.735387	30.341203	Type_S
109	36.686353	5.010884	41.948751	31.675469	84.241415	0.664437	Type_H
30	41.646916	8.835549	36.031975	32.811367	116.555168	-6.054538	Nrmal
257	44.216464	1.507075	46.110339	42.709390	108.629567	42.810481	Type_S
93	34.382299	2.062683	32.390820	32.319617	128.300199	-3.365516	Normal
116	63.834982	20.362507	54.552434	43.472475	112.309492	-0.622527	Type_H
14	89.834676	22.639217	90.563461	67.195460	100.501192	3.040973	Nrmal
297	58.828379	37.577873	125.742385	21.250506	135.629418	117.314683	tp_s
207	78.425951	33.425951	76.277439	45.000000	138.554111	77.155172	Type_S
261	67.263149	7.194661	51.696887	60.068488	97.801085	42.136943	Type_S
113	53.572170	20.460828	33.100000	33.111342	110.966698	7.044803	Type_H
208	63.172987	6.330911	63.000000	56.842076	110.644021	42.608076	Type_S

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree
count	310.000000	310.000000	310.000000	310.000000	310.000000	310.000000
mean	60.496653	17.542822	51.930930	42.953831	117.920655	26.296694
std	17.236520	10.008330	18.554064	13.423102	13.317377	37.559027
min	26.147921	-6.554948	14.000000	13.366931	70.082575	-11.058179
25%	46.430294	10.667069	37.000000	33.347122	110.709196	1.603727
50%	58.691038	16.357689	49.562398	42.404912	118.268178	11.767934
75%	72.877696	22.120395	63.000000	52.695888	125.467674	41.287352
max	129.834041	49.431864	125.742385	121.429566	163.071041	418.543082

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree
P_incidence	1.000000	0.629199	0.717282	0.814960	-0.247467	0.638743
P_tilt	0.629199	1.000000	0.432764	0.062345	0.032668	0.397862
L_angle	0.717282	0.432764	1.000000	0.598387	-0.080344	0.533667
S_slope	0.814960	0.062345	0.598387	1.000000	-0.342128	0.523557
P_radius	-0.247467	0.032668	-0.080344	-0.342128	1.000000	-0.026065
S_Degree	0.638743	0.397862	0.533667	0.523557	-0.026065	1.000000

	Class	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree
0	Normal	51.685244	12.821414	43.542605	38.863830	123.890834	2.186572
1	Type_H	47.638407	17.398795	35.463524	30.239612	116.474968	2.480251
2	Type_S	71.514224	20.748038	64.110108	50.766186	114.518810	51.896687

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree
count	310.000000	310.000000	310.000000	310.000000	310.000000	310.000000
mean	59.893743	16.060995	51.689825	42.697607	118.335308	21.474538
std	16.139975	7.397731	18.071145	12.656481	10.515370	24.167936
min	26.147921	-2.970024	14.000000	13.366931	93.692209	-11.058179
25%	46.430294	10.839352	37.000000	33.347122	111.974906	1.603727
50%	58.691038	16.000000	49.562398	42.349084	118.000000	11.767934
75%	72.313279	20.338937	62.964777	52.475365	125.127803	37.277675
max	96.657315	34.310989	100.744220	79.695154	143.560690	88.301486

Project 1¶

DOMAIN: Healthcare¶

Import and warehouse data:¶

Data cleansing:¶

Data analysis & visualisation:¶

Lets check the distribution of data for all feature variables, we can use distplot, boxplot/violinplot, swarm plot for univariate and bivariate analysis.¶

Observation¶

Data pre-processing:¶

Model training, testing and tuning:¶

Conclusion and improvisation:¶

Project 2¶

Data analysis & visualisation:¶

Response variable distribution check¶

Data pre-processing:¶

Model training, testing and tuning:¶

NB classifier¶

Conclusion and improvisation:¶

Improvements¶

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree
0	38.505273	16.964297	35.112814	21.540976	127.632875	7.986683
1	54.920858	18.968430	51.601455	35.952428	125.846646	2.001642
2	44.362490	8.945435	46.902096	35.417055	129.220682	4.994195
3	48.318931	17.452121	48.000000	30.866809	128.980308	-0.910941
4	45.701789	10.659859	42.577846	35.041929	130.178314	-3.388910
...	...	...	...	...	...	...
305	80.111572	33.942432	85.101608	46.169139	125.593624	21.000000
306	95.480229	16.000000	59.000000	48.930176	96.683903	77.283072
307	74.094731	18.823727	76.032156	55.271004	128.405731	73.388216
308	87.679087	20.365613	93.822416	67.313473	120.944829	76.730629
309	48.259920	16.417462	36.329137	31.842457	94.882336	28.343799

	ID	Age	CustomerSince	HighestSpend	ZipCode	HiddenScore	MonthlyAverageSpend	Level	Mortgage	InternetBanking	CreditCard	LoanOnCard
4642	4643	65	40	143	95616	4	6.60	2	0	1	0	1.00
1897	1898	54	29	98	93065	1	0.10	3	0	0	0	0.00
4839	4840	34	8	52	95060	4	0.20	1	0	1	0	0.00
2465	2466	58	34	25	90059	2	0.30	1	123	1	0	0.00
3613	3614	35	11	148	92672	1	5.80	3	0	1	0	1.00
4698	4699	48	22	162	94143	3	1.40	1	400	0	0	1.00
2650	2651	64	40	52	95060	2	1.10	1	226	0	0	0.00
4439	4440	33	7	104	94542	2	3.60	3	0	0	1	1.00
467	468	45	20	39	90232	1	2.40	1	0	0	1	0.00
3798	3799	55	25	35	93943	3	1.00	3	144	1	1	0.00

	ID	Age	CustomerSince	HighestSpend	ZipCode	HiddenScore	MonthlyAverageSpend	Level	Mortgage	Security	FixedDepositAccount	InternetBanking	CreditCard	LoanOnCard
count	5000.00	5000.00	5000.00	5000.00	5000.00	5000.00	5000.00	5000.00	5000.00	5000.00	5000.00	5000.00	5000.00	4980.00
mean	2500.50	45.34	20.10	73.77	93152.50	2.40	1.94	1.88	56.50	0.10	0.06	0.60	0.29	0.10
std	1443.52	11.46	11.47	46.03	2121.85	1.15	1.75	0.84	101.71	0.31	0.24	0.49	0.46	0.30
min	1.00	23.00	-3.00	8.00	9307.00	1.00	0.00	1.00	0.00	0.00	0.00	0.00	0.00	0.00
25%	1250.75	35.00	10.00	39.00	91911.00	1.00	0.70	1.00	0.00	0.00	0.00	0.00	0.00	0.00
50%	2500.50	45.00	20.00	64.00	93437.00	2.00	1.50	2.00	0.00	0.00	0.00	1.00	0.00	0.00
75%	3750.25	55.00	30.00	98.00	94608.00	3.00	2.50	3.00	101.00	0.00	0.00	1.00	1.00	0.00
max	5000.00	67.00	43.00	224.00	96651.00	4.00	10.00	3.00	635.00	1.00	1.00	1.00	1.00	1.00

	ID	Age	CustomerSince	HighestSpend	ZipCode	MonthlyAverageSpend	Mortgage
ID	1.00	-0.01	-0.01	-0.02	0.01	-0.03	-0.02
Age	-0.01	1.00	0.99	-0.05	-0.03	-0.05	-0.01
CustomerSince	-0.01	0.99	1.00	-0.05	-0.03	-0.05	-0.01
HighestSpend	-0.02	-0.05	-0.05	1.00	-0.02	0.65	0.21
ZipCode	0.01	-0.03	-0.03	-0.02	1.00	-0.00	0.01
MonthlyAverageSpend	-0.03	-0.05	-0.05	0.65	-0.00	1.00	0.11
Mortgage	-0.02	-0.01	-0.01	0.21	0.01	0.11	1.00

	LoanOnCard	Age	CustomerSince	HighestSpend	HiddenScore	MonthlyAverageSpend	Level	Mortgage
0	0.00	45.38	20.15	66.29	2.37	1.73	1.84	51.87
1	1.00	45.07	19.84	144.75	2.61	3.91	2.23	100.85